A Kalman Filter based Sensor Fusion Technique for Balancing a 2-Wheel System

The objective of the presented work is to design and implement a sensor fusion algorithm using Kalman Filter for balancing the system on 2-wheels. The system uses inertial sensors such as 3-axis linear accelerometer and dual-axis gyroscope to calculate the tilt for balancing. The estimation algorithm i.e. Kalman filter continuously and recursively corrects the values obtained by mathematical integration of the velocity, measured using gyroscope at the rate of 20Hz. The correction is performed using the inclination value obtained from accelerometer. This reduces the integration drift that originates from errors in the angular velocity signal. In addition, the gyroscope offset is continuously calibrated. The tilt estimated by the Kalman filter is given to PID algorithm with a reference of 0 radian, to balance the system. The result shows the need of Kalman filter to remove sensor noise. The control and filter algorithm are implemented on Atmega32 microcontroller. This study reinforces the significance of sensor fusion for optimum performance. The conception presented in this paper will be of assistance in existing applications and in new designs.